Table 1.

#	Author/Year	Country	Methodological approach	Participants	Research questions/aims	Main themes
1	Aidoo and Harpham, 2001	Zambia	Qualitative interviews using modified version of Kleinman's EM questions; group interviews with HCWs	Household survey in Mtendere, Lusaka (n = 323) identifying 139 women with depression (scoring >7 on SRQ) 10 HCWs providing MH services to the Mtendere community- 6 x psychiatrists, 1 x psychologist and 3 x social workers (urban Lusaka).	To compare the EMs of HCWs and low income urban women from the community in which they work	Name given to women's experience; cause of stress; onset of stress; how the experiences of stress make women feel; severity of stress; women's greatest fears about their experience of stress; choice of treatment; influences leading to choice of treatment; results hoped to be received from treatment
2	Andersen et al. (2015)	South Africa	Informed by grounded theory, in-depth interviews, open coding followed by axial and selective coding	14 adult attendees of an infectious disease clinic in a township east of Cape Town with a diagnosis of MDD (using the MINI)	To describe the experience of black South Africans living with HIV and depression	Symptoms: affective, cognitive, behavioural; relevance of HIV to depression symptoms; depression treatment history
3	Bolton (2001)	Rwanda	Ethnographic methods: a) free-listing of MH symptoms and disorders; b) informant interviews; c) pile sorts to confirm relationships between symptoms/disorders	Residents of 2 x rural administrative regions near Kigali: a) 41 people judged to be knowledgeable about community problems by interviewers; b) people identified in a) as individuals people would consult about MH problems; c) convenience sample (n = 40)	To investigate how Rwandans perceive the mental health effects of the 194 genocide; to investigate the local validity of western mental illness concepts; to provide data to adapt existing MH assessments for local use	Guhahamuka and agahinda include all the symptoms categories required for DSM-IV diagnoses of PTSD and depression
4	Burgess and Campbell, 2014	South Africa	In-depth interviews, grounded thematic analysis	19 women, living in Umkhankude sub-district, in North Kwa Zulu Natal province. Randomly selected from list of participants in a community survey about social deprivation and HIV. Participants completed CES-D (depression), Beck Anxiety Inventory, Harvard Trauma Questionnaire but were not selected on basis of scores.	To obtain a fine-grained account from women about perceived links between HIV and MH; to explore women’s understanding of mental distress among a non-clinical population; to understand women’s’ efforts to cope	Women were economically dependent upon men: their distress was often grounded in abandonment and impoverishment. Conflict in other relationships, HIV and violence were also important drivers of distress. Coping often involved psychological reframing of life struggles towards acceptance. Government grants enabled women to survive and income generation opportunities were sought. These strategies rarely led to long-term resolution.
5	Familiar et al. (2013)	Burundi	a)Free-listing and b) semi-structured interviews with key informants	Residents in 5 communities in the Kibuye Health District (high population density area) purposive sampling: a) n = 38; b) n = 23	To explore community perceptions of mental distress, understand key concepts and associated behaviours to inform future service delivery and policy development	Ukutiyemera corresponds to DSM-IV description of dysthemia and depressive disorder. Traditional healer, priests, community health centres were all potential sources of help. However, HCWs reported that many remain at home without treatment. Most important impact of mental distress was loss of assets.
6	Irankunda and Heatherington, 2017	Burundi	Mixed methods: semi-structured interviews- asking about treatment preference for three conditions (quantitative) and rationale for treatment expectations (qualitative). Constant comparison involving three researchers	N = 198 attendees of primary care at the Village Health Works clinic, Kigutu (service run by NGO- no mental health screening as no mental health services available; rural)	To describe expectations of efficacy of four different treatments (spiritual, traditional, medication, evidence-based psychosocial interventions) across three key syndromes, including akabonge (depression-like symptoms)	Explanations for why the four treatment options would/would not work
7	Irankunda and Heatherington (2017)	Burundi	a)pilot study- open ended group discussions; b) semi-structured interviews asking about three syndromes identified in a) and symptoms/causes of these; c) semi-structured interviews presented syndromes identified in b) and asking what they would be called	a) n = 761 (14 groups, ranging from 25 to 50 participants); b) n = 542 (52% men); c) n = 143 (46% men). Attendees of the Village Health Works clinic, Kigutu (service run by NGO- no mental health screening as no mental health services available; rural). a) was conducted as part of the public health discussions held with people in the waiting room each morning (education pour la sante).	To derive a basic taxonomy and description about mental health problems.	Evidence of three major conditions: the closely related conditions of kuyinga and akabonge/ibonge; guhahamuka and gusara. Partial overlap with depression, PTSD and schizophrenia, respectively. Causes of akabonge include: war, problems, loss of loved ones, sadness, loneliness
8	Johnson et al. (2009)	Uganda	Semi-structured interviews based on Kleinman's EM questions	Residents of Kampala or surrounding peri-urban districts: community members (n = 135); HCWs (n = 111) (from 44 districts- recruited from markets, businesses, health services).	To compare EMs between community members and the HCWs who provide a service in that community	Contextual relevance of depression; labelling and conceptualisation; aetiology; impact and social meanings; help-seeking; type of treatment; treatment expectations; difference based on social characteristics
9	Johnson et al., 2017a	Uganda	Semi-structured interviews based on Kleinman's EM questions	Residents of Kampala or surrounding peri-urban districts: community members (n = 135); HCWs (n = 111) (from 44 districts- recruited from markets, businesses, health services); people meeting criteria for depression and seeking therapy services (n = 33, from 17 districts)	To explore whether EMs predicted help-seeking through assessment of the relationship between problem conceptualisation and treatment	Problem conceptualisation was not a predictor of treatment choice among either community members, HCWs or the sample of people with depression.
10	Johnson et al., 2017a,b	Uganda	Semi-structured interviews based on Kleinman's EM questions	Adult patients at traditional healing and psychiatry clinics (n = 30) and patient-provider dyads (n = 8) near Kampala, Uganda	To investigate differences in EMs associated with help-seeking (traditional or psychiatric services) among patients and providers	Patients in both settings had similar EMs in terms of symptoms, perceived cause, seriousness, impact of depression. However, those attending traditional clinics had a preference for herbal treatments, whilst patients at psychiatric clinics were more likely to desire biomedical treatment.
11	Kidia et al. (2015)	Zimbabwe	In-depth interviews	Purposive sample of attendees at the HIV clinic of Parirenyatwa Hospital, Harare (one of the largest treatment facilities in the country) who were living with HIV and scored >5 on the Shona version of the SRQ-8 (n = 47)	To explore lived experience of adults living with HIV and co-morbid CMD with poor adherence to ART in order to develop a culturally appropriate intervention	Challenges- poverty, stigma, marital problems, symptoms of CMDs; impact of challenges on adherence and access to ART: poverty, stigma; intervention ideas: family engagement and disclosure, income generation and transport, privacy
12	Monteiro & Balogun 2014	Ethiopia	Semi-structured interviews	Purposive sampling of HCWs (n = 35), lay people (n = 75) and traditional healers (n = 5) resident in Addis Ababa (n = 82) and Asella, rural Southern Ethiopia (n = 33).	To explore ideas about the definition and expression, causation and treatment of mental illness and perceptions of depression, anxiety and psychosis? To examine whether community members/laypersons, healthcare workers and traditional healers differ in their attitudes, beliefs and practices regarding mental illness	Participants from all three groups agreed that you would know someone had depression from their negative affect/emotions; the most common cause of depression identified was loss of loved one/death. HCWs and lay people felt that advice and counselling and social support were the most appropriate treatment; where as traditional healers thought opted for traditional/cultural treatments
13	Murray et al. (2017)	Uganda	Semi-structured interviews	Purposive sample (n = 9), those scoring > mean on Hopkins Symptom Checklist- caregivers of young children living with HIV, participants in the control arm of an RCT (n = 60) of parenting intervention- attendees of local clinics/NGOs providing HIV care in Tororo and Busia district, Eastern Uganda (rural).	To create a contextually embedded conceptual framework of the relationship between caregiver mental health and HIV-infected child well-being, in order to inform support services for families living with HIV	Fulfilling the caregiving role; how caregiver mental health affects children; how child sickness affects caregivers' lives; mental health and inability to provide; duality of support and isolation
14	Myers et al. (2018)	South Africa	Qualitative interviews	Maximum variation sampling of patients attending 17 Primary Healthcare Clinics receiving care for diabetes (n = 11) or HIV (n = 19) in the Western Cape (n = 30, includes n = 1 co-morbid diabetes and HIV); scoring >16 on the CES-D or >8 < 22 on the AUDIT (for Alcohol Use Disorder). “Almost all” scored >16 on CES-D.	To describe patients living with chronic disease perceptions of acceptability of mental health counselling in the context of PHC as well as preferences for mode of delivery of counselling	Screening for mental health problems was felt to be useful as poor awareness meant people wouldn't proactively seek help. Coping with stress, often linked to chronic disease diagnosis was a key problem/need. Brief counselling was preferred, delivered by specialist mental health counsellor
15	Okello and Ekblad, 2006	Uganda	a) focus group discussion, c) in-depth interviews with key informants using case vignettes. Constructivist version of grounded theory was selected as frame of analysis	Community members from Bajjo, small village in Mukuno (23 km from Kampala, semi-rural). a) (n = 5) traditional and faith healers; b) (n = 25, of which n = 13 women) 4 groups: secondary school girls; women (mean age 35yrs); men (mean age = 38yrs); primary school teachers (mean age = 35yrs); c)	To assess the feasibility of using case vignettes to explore local explanatory models for various sub-types of depressive illness- including aetiological factors, perceived effects of depressive symptoms and appropriate forms of help	Identity given to symptoms according to type of depression; aetiological factors associated with onset of depression- psychological factors, socioeconomic, spiritual/cultural factors, biological/physical factors; effects of depressive symptoms; sources of care for depression
16	Okello and Neema, 2007	Uganda	Semi-structured interviews based on Kleinman's EM questions. Analysis was informed by symbolic interaction theory- “all objects, events, situations acquire their meaning through a process of human interpretation”	Purposive sample of people with an Axis I depression diagnosis accessing a mental health clinic at Mulago Hospital, Kampala (n = 22) (national referral hospital, regional hospital for the central region)- referred to researchers by psychiatrists working at the clinic. Clinician diagnosis confirmed by researcher using MINI.	To explore how people diagnosed with depression conceptualised their illness and how the conceptualisation shaped courses of action in the search for help. Intention was to capture the complexity of the decision to seek help	Somatization, social meaning of depression and help-seeking; meaning and perceived consequences of illness; “How did I get here”- making sense of psychiatric admission; variations in the causal attribution and the role of significant others in help-seeking
17	Okello et al. (2012)	Uganda	Semi-structured interviews, thematic analysis	Consecutive attendees receiving ART at a large HIV treatment centre (caseload n = 11000), diagnosed with major depressive disorder by clinic psychiatrist (DSM IV) (n = 26, n = 11 had received antidepressants treatment, n = 15 interviewed before starting antidepressants)	To explore how depressed people living with HIV in Uganda conceptualised and described their depression and its manifestation in the context of ART and anti-depressant treatment	Depression symptoms reported by participants; attributions of depression in the context of ART; the effects of ART alone on the psychological health of clinically depressed individuals; the effects of antidepressants on the psychological wellbeing of clinically depressed ART patients
18	Nakimuli-Mpungu et al., 2014	Uganda	Focus groups, thematic analysis	Attendees of Peter C. Alderman Foundation (PCAF) trauma clinics in Gulu and Kitgum, northern Uganda (areas that have experienced >20 years of civil war) who were receiving treatment for depression/had experienced depression in the past (identified from clinic records by staff) and their caregivers (n = 110).	To obtain information on the cultural understanding of depression symptoms, complications and treatment methods used in post-conflict communities in northern Uganda, in order to inform the development of an indigenous group support intervention to treat depression	Community perceptions of depression and mental health problems; community strategies used to combat depression in the acholi community; community perceptions of counselling; structure and content of group support intervention;
19	Patel et al. (1995)	Zimbabwe	Focus groups	N = 9 Focus Group Discussions (FGDs) in total- participants were selected by investigators, who had existing links with the relevant stakeholder groups: n = 30 village community workers took part in n = 3 FGDs; n = 22 traditional and faith healers took part in n = 3 FGDs; n = 9 community psychiatric nurses took part in a FGD and n = 15 relatives of patients attending out-patient clinics at the Parirenyatwa Hospital, Harare took part in n = 2 FGDs	To generate information on the concepts of mental illness from a range of care-providers in Harare	Symptoms; impact; sources of care
20	Patel et al. (1995)	Zimbabwe	Semi-structured interviews based on EMIC followed by CISR	N = 109 attendees of primary care at three health centres and the clinics of four traditional healers in high density suburbs of Harare	To elucidate symptom profiles and describe explanatory models of illness among people considered to have mental illness by care providers	Somatic complaints were the most common presentation but only one fifth of patients thought that their illness was purely one of the body. Illness was chronic- this is likely to reflect bias of providers. Spiritual causes were the most frequently mentioned explanation, with “kufungisa”- thinking too much identified as a proximate cause.
21	Psaki and Hindin, 2016	Ghana	Mixed methods to examine the validity of the CES-D scale, nested in Family Health and Wealth Study (FHWS- multi-country cohort study) including: a) focus groups and b) In-depth interviews; inductive content analysis	a) n = 12 FGDs, with a total of n = 95 participants (42 men and 53 women) who were FHWS participants; b) n = 19 religious leaders, healthcare providers, community elders/leaders (11 men, 8 women, not FHWS participants)	To use mixed methods to assess whether the CES-D 10 scale effectively captures culturally relevant domains of depressive symptoms among men and women of reproductive age in Kumasi, Ghana.	Most common symptoms associated with depression were: loss of concentration, crying, loss of appetite, becoming quiet or withdrawn. Suicide was a common impact of depression
22	Petersen et al. (2013)	South Africa	In-depth interviews based on Kleinman's EMs approach. Thematic content analysis- a priori plus emerging codes, constant comparison	N = 35 women living with HIV and depression, attendees of HIV/AIDS two clinics in KwaZulu-Natal and North-West Province (peri-urban). Participants were those who attended either clinic on 4 randomly selected days who scored >8 on the SRQ-20 and were diagnosed with major depressive disorder using the Structured Clinical Interview for DSM-IV Diagnosis (SCID). Excluded if pregnant/delivered baby in last 6 months.	To understand the context and local understandings of depression among women living with HIV, with a view to informing the content of a culturally acceptable care package for depression; to develop an understanding of how best to deliver this package using a collaborative care task-sharing approach within existing resource constraints	Perceived causes of depressive symptoms and exacerbating factors: being HIV-positive, stigma & discrimination, lack of social support, partner rejection/abuse/abandonment, other family problems, poverty related stress, trauma and loss; depressive symptoms; current coping strategies and possible interventions
23	Sorsdahl et al., 2010	South Africa	Focus groups using vignettes from the Short Explanatory Model Interview (SEMI); in-depth interviews. Framework approach to analysis	Convenience sample of n = 50 (4 FGDs with 8 healers in each group and 18 in-depth interviews). Traditional healers were selected from those who attended a workshop conducted by the South African Depression and Anxiety Group (mental health advocacy group) in Mpumalanga.	To identify concepts, causes, and treatments for mental disorders among traditional healers, including exploring basic concepts and contrasting responses to a psychotic vignette with responses to vignettes representing non-psychotic mental disorders	Healers did not think case described in depression vignette was a mental illness but was a problem caused by psychological reasons. Those who thought the patient was suffering from an illness (68%) believed this required attention of a traditional healer (40%) or a western doctor (28%).
24	Ventevogel et al. (2013)	Burundi, Democratic Republic of Congo, South Sudan	“Rapid ethnographic assessment”- Focus Group Discussions (FGDs) and key informant (KI) interviews. FGD ad KI participants were asked to talk about “problems or illness that manifest through problems in thinking, feeling or behaving”. KI participants were also asked to talk about their work. Thematic analysis.	n = 31 FGDs with a total of n = 251 participants, across four broadly agriculture-focussed rural sites affected by conflict: Kwajena Payam and Yei (South Sudan); Butembo (DRC); Kibuye (Burundi). Separate FGDs were held for men and women. The aim was to achieve sample characteristics broadly consistent with site demographics. N = 26 KI interviews were held with local experts on mental health problems- traditional/religious healers, healthcare workers and policy makers	To explore local concepts of mental disorder in four settings in Africa	Identification of five syndromes with similarities to depression/anxiety (South Sudan- yeyeesi. “people whose mind is busy with thoughts”, DRC- Amutwe alluhire “tired head”, Burundi- ibonge/akabonge- dwelling on loss, South Sudan- nger yer “cramped stomach, always sad”. Causes were perceived to be loss/worry- including of livelihood and properties but particularly loved ones. Social interventions to provide support and to address isolation were recommended by participants.
25	Wilk and Bolton, 2002	Uganda	Free-listing and key informant interviews	Free-listing participants were 50 local people from 10 villages. Participants were asked “what are the main problems that affect the people of this community as a result of HIV?“. Twenty key informants (people whom local consultants identified as being knowledgeable about mental health). Interviews were focussed on problems identified in free-listing.	To understand how people perceive the mental health effects of HIV, including examining the validity of western concepts of PTSD and depression.	Free-listing identified eight psychological problems from a total of 30 problems reported. Two syndromes were identified “Yo'kwekyawa”- “hating oneself” and “Okwekubaziga”- “pitying oneself”, which, between them, include all of the nine major DSM-IV depression criteria.